Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Author for correspondence: J. Díez Tel: +33 383 39 41 45 Fax: +33 383 39 40 69 Email:email@example.com
• Genetic diversity and host specificity of Pisolithus is reported here in exotic (Eucalyptus) and native hosts in the western Mediterranean region.
• Polymorphism in the internal transcribed spacer (ITS) sequences of the nuclear rDNA of Pisolithus was analysed. Sequences for 17 isolates associated with native Mediterranean hosts and Eucalyptus were compared with those in the GenBank DNA database using distance and parsimony methods.
• Bootstrap analysis showed clustering of all Pisolithus isolates associated with Mediterranean hosts. The ITS sequences suggest the occurrence of several ecological species adapted to exploit different soil types (basic, acid and clayey slate-derived soils), with specificity for particular indigenous hosts. Isolates from eucalypt plantations in Brazil, Kenya and the Mediterranean grouped together with eucalypt-associated Australian isolates. Transfer to native hosts did not occur; the host specificity range of these exotic strains might prevent out-competition and interbreeding with local species.
• Pisolithus spp. in eucalypt plantations in the Mediterranean basin are of Australian origin; the co-introduction of the ectomycorrhizal fungi might explain the success of these exotic forest plantations.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The Mediterranean region forms a belt around the Mediterranean Sea in Southern Europe, the Levant and North Africa. The western region comprises the Iberian Peninsula and some areas of Morocco. The natural vegetation corresponds to forests of sclerophyllous oaks, adapted to a climate regime characterized by a dry summer and a cool moist winter (Archibold, 1995). In these regions, efficient pine germination after fire and extensive plantations account for today’s wide-spread pine woodlands. On siliceous soils, dense scrub formations, known as jaral (Cistus spp.), often dominate (Schultz, 1995). During the last two centuries extensive plantations of Eucalyptus have been established in the Mediterranean region to produce paper pulp. Such eucalypt species are of Australasian origin, and thus exotic to the Mediterranean region.
Most Mediterranean forest trees are obligate ectomycorrhizal plants, establishing mycorrhizal symbiosis with ascomycetes and basidiomycetes. Pisolithus is a common ectomycorrhizal fungus in native Quercus forests, Pinus woodlands and Cistus scrubs (Calonge & Demoulin, 1975). This fungus also occurs in plantations of eucalypt. Pisolithus has been reported on acid (siliceous or slate-derived) and basic (calcareous and marsh-gypsum) soils. Pisolithus is often regarded as a cosmopolitan ectomycorrhizal fungus with a wide host range, establishing mycorrhizas with angiosperms and gymnosperms (Marx, 1977). Moreover, Pisolithus isolates are commonly used as inoculants to enhance tree establishment and growth of pine and eucalypt plantations worldwide (Garbaye et al., 1988).
A great variation in the effects of inoculation with different strains of Pisolithus on forest trees has been reported (Burgess et al., 1994). Since fungal variability plays a major role in strain selection for mycorrhizal inoculation programmes, some work has been done on Pisolithus heterogeneity. There is considerable polymorphism in terms of carpophore, spore and culture morphology among Pisolithus strains. Large variations in colony growth rates, enzyme activity, polypeptide patterns and mycorrhizal ability have been reported (Ho, 1987; Kope & Fortin, 1990; Lamhamedi et al., 1990; Burgess et al., 1994, 1995).
Since Pisolithus was described, several taxa have been proposed based on distinctive carpophore and basidiospore morphology (Marx, 1977). Morphological differences were considered as nondiagnostic, and taxa within the genus Pisolithus were regarded as conspecific (Coker & Couch, 1928; Pilát, 1958). However, several new species have been described, that is Pisolithus microcarpus Coke & Mass, P. kisslingi E. Fisch, P. pussillum Pat, and P. auriantioscabrosum (Watling et al., 1995).
Information on the genetic polymorphism of Pisolithus is scattered in different locations, and is sometimes contradictory and hard to reconcile. For instance, in North America, Grand (1976) found no correlation between the basidiocarps and basidiospores of Pisolithus, their geographical location, habitat or plant association. However, others studies have stressed that host and world location may play a significant role in such variability (Marx, 1981; Malajczuk et al., 1990; Burgess et al., 1994). Burgess and co-workers (1995) reported a correlation between geographical origin, basidiospore morphology and polypeptide patterns in Australian isolates. In addition, host specificity among Pisolithus isolates has been revealed in mycorrhizal synthesis experiments (Malajczuk et al., 1990; Burgess et al., 1994).
Based on basidiospore morphology and incompatibility mating tests, Kope & Fortin (1990) proposed that the genus Pisolithus comprises several biological species. DNA-based methods have provided further support for this hypothesis (Cairney et al., 1999). Junghans et al. (1998) reported that RAPD analysis (Random Amplification of Polymorphic DNA) clustered isolates in two main groups according to their host and geographical origin. Gomes et al. (1999) confirmed these data on the same isolates by Restriction Fragment Length Polymorphism (RFLP) in the internal transcribed spacer (ITS) regions amplified by Polymerase Chain Reaction (PCR) and Gomes et al. (2000) on mitochondrial DNA. Farmer & Sylvia (1998) based on RFLP analysis of the ITS sequences from several isolates of Pisolithus, suggested that this taxon represents a species complex. Furthermore, Anderson et al. (1998) and Martin et al. (1998), after studying ITS sequences of Pisolithus isolates from a defined region in New South Wales (Australia) and in Kenya, respectively, suggested that Pisolithus may comprise several species.
In the Western Mediterranean region, it is envisaged that mycorrhizal inoculation with selected isolates of Pisolithus may become a common practice in eucalypt plantations, in order to improve seedling survival and growth after outplanting. Since mycorrhizal fungi of Australian origin may out-compete native fungal symbionts or even interbreed with native strains, environmental concerns arise about the introduction of symbiotic microorganisms together with these exotic trees. Little information on the behaviour of Pisolithus strains in natural ecosystems is available, particularly when plantations of exotic Pisolithus hosts co-exist with native forests (Martin et al., 1998). If natural ectomycorrhizal inocula are absent and mycorrhizal inoculation is not used, exotic forest plantations usually fail. However, most eucalypt plantations in Mediterranean regions have been successful without undertaking artificial inoculation. Two hypotheses can be proposed: either exotic ectomycorrhizal fungi were co-introduced with eucalypt seedlings, or some compatible Mediterranean ectomycorrhizal fungi were able to colonize seedlings of Eucalyptus. The aim of this work was to study the genetic variability of Pisolithus in the Mediterranean region and its correlation with ecological and host factors. In addition, it aimed to discover whether Pisolithus strains in eucalypt plantations are native or were introduced with eucalypt seedlings. Sequencing of the ITS regions from 17 isolates associated to Mediterranean hosts and Eucalyptus was conducted. These sequences were compared with those available in DNA database by using distance and parsimony methods. The ITS sequence analysis approach was chosen because its has previously been shown to be useful in identifying and grouping Pisolithus strains (Anderson et al., 1998; Martin et al., 1998).
Materials and Methods
Sampling and source of collections
Carpophores of Pisolithus were collected from native forests of sclerophyllous oaks, pine woods, Cistus shrublands and plantations of Eucalyptus (Table 1). One hundred and twenty-five samples collected between 1994 and 1998 were studied. Specimens associated with Afzeliaquanzensis Welw and Pinus and eucalypt plantations in Kenya were also studied (specimens K-915, esp07, K354, respectively). Collections were grouped in five major groups according to their morphological similarities. Fifteen samples were chosen covering all Pisolithus morphotypes to study the polymorphism in the ITS/5.8S sequences of the nuclear rDNA. Three major soil types were considered, clayey slate-derived and sandy-siliceous soils (both acid), and basic soils (limey or marl-gypsum). Samples were chosen from both the Iberian Peninsula and Morocco. In addition, isolates from the lowland woodlands of Kenya associated with Afzelia quanzensis and Pinus caribaea Mor. were sequenced (Tables 2 and 3).
Table 1. Collection sites of Pisolithus basidiocarps in the Western Mediterranean region
Table 3. List of Pisolithus isolates whose internal transcribed spacer (ITS) sequences were obtained in the present work. DNA used in polymerase chain reaction (PCR) was isolated from mycelium (*) or fruitbodies (**).
For each Pisolithus morphotype, carpophore features were recorded, including size, shape and colour. Spores were studied by light microscopy. Voucher specimens were deposited at the Alcalá University Herbarium, Spain.
Fungal isolation and growth conditions
Isolates were obtained from selected carpophores on Melin–Norknans medium agar medium (Marx, 1969). All cultures were maintained on MMN agar medium at 25°C in the dark. Studied isolates were deposited at the Collections of Ectomycorrhizal Fungi of the Plant Biology Departments of the Universities of Alcalá and Murcia.
DNA extraction and ITS amplification
Samples for DNA analysis were excised from the outer part of a mycelial colony for isolates referred to as * in Table 3. Mycelial samples were preserved at −20°C until DNA extraction. For those isolates referred to as **, a piece was excised from the inner part of the basidiocarp to avoid contamination by other microorganisms. Approx. 20–50 mg of samples were used for each DNA extraction, performed using the cetyl-trimethyl-ammonium bromide (CTAB) protocol according to Gardes et al. (1991). The ITS regions of nuclear rDNA were amplified in the same way as Anderson et al. (1998), with the ITS1F and ITS4B primers (Gardes & Bruns, 1993). The ITS4B primer (specific for basidiomycetes) was used to avoid nonspecific amplification of ascomycete mould and yeast contaminants. Amplifications were performed on a Mastercycle Personal PCR system (Eppendorf, Hamburg, Germany). Controls with no DNA were included in every set of amplifications to test for the presence of DNA contamination of reagents and reaction buffers.
Amplified DNA was purified with GeneClean Kit (Bio101, Carlsbad, CA, USA). The amplification products were size-fractionated using 1.5% agarose gels in a 1× Tris-borate-EDTA buffer (Sambrook et al., 1989). The gels were stained with ethidium bromide and photographed under ultraviolet light. λ phage DNA digested with HindIII was included in the gels as molecular size marker. DNA concentrations were quantified onto agarose gels with a Low DNA Mass ladder (GibcoBRL, Life technologies, Cergy Pontoise, France), and sizes calculated by using a 100-bp (base pair) ladder (GibcoBRL).
Sequencing of the amplified ITS regions
Sequencing reactions were performed directly on purified PCR products with ITS1 and ITS4 primers (White et al., 1990). Both strands were sequenced with the Taq DyeDeoxy Terminator Cycle Sequencing Kit (Perkin Elmer, Norwalk, CT, USA). The sequence products were analysed using an ABI model 373 DNA fluorescent sequencer (Perkin Elmer). The obtained sequences were deposited in the National Center for Biotechnology Information (NCBI) GenBank (Table 3).
The search for sequence identity in the GenBank DNA database was conducted by Gapped Blast (NCBI) (Altschul et al., 1997). This search found the Pisolithus ITS sequences which can be seen in Table 2. Obtained generated sequences were aligned with the ITS Pisolithus sequences found in GenBank using MultAlin (Corpet, 1988). The alignment was deposited in TreeBASE (http://www.herbaria.harvard.edu/treebase/). Boundaries of the regions were determined through comparison with known sequences of rDNA genes and spacers.
The aligned sequences were analysed by distance and parsimony methods. Distances were calculated according to the Junkes and Cantor’s model, as:
(DAB, the distance between sequences A and B;Su, the number of positions showing a substitution; G, the number of gaps in one sequence with respect to the other; I, the number of identical nucleotides; and T, sum of I, Su and G.) Insertions and deletions were taken into account. Pairwise distance analysis were performed using this index. Tree topology was inferred using the Neighbour-Joining (NJ) method (Saitou & Nei, 1987). The Bootstrap method (Felsenstein, 1985) was performed to evaluate the reliability of the obtained tree topologies, using 1000 bootstrap replications. Distance analysis and tree drawing were carried out using the TreeCon program (Van de Peer & De Wachter, 1993).
Parsimony trees were inferred through the heuristic method, with the aid of PAUP 3.1.1 (Swofford, 1993). Validity of clades was tested using bootstrap analysis (Felsenstein, 1985). The analysis was conducted using 1000 replications, retaining clades compatible with the 50% majority-rule in the bootstrap consensus tree using the Nearest-Neighbour Interchange (NNI) branch-swapping option, and saving no more than 100 trees with < 700 length each replication. The 50% majority rule consensus tree is available on-line at TreeBASE. The parsimony tree of Fig. 1 was drawn with the aid of TreeView program (Page, 1996).
Four distinct morphotypes were defined on the basis of the spore and carpophore morphologies: three Mediterranean host-associated morphotypes and one Eucalyptus-associated morphotype (Eucalyptus type I) (Table 4). A number of fruit bodies presented intermediate features and were therefore not taken into account in the present work (probably immature fruit bodies). The morphotypes associated with Mediterranean hosts (i.e. oak, pine and Cistus) were referred as slate, acid and basic types.
Table 4. Basidiospore and carpophore characteristics of Pisolithus morphotypes
Main basidicarp feature
Spore mass colour
Spore ornamentation and size
Numbers in brackets indicate maximum size.
Bright yellow mylecum at the base
Short spines 7.5–9(11) µm
Large, often > 12 cm
long and with a massive speudostump
Pseudotump with basal thick rhizomorphs
Short spines 7.0–8.5(10) µm
Eucalyptus type I
Large, often > 12 cm long
Long spines 6–7.5(10) µm
Eucalyptus type II
Small, often < 12 cm long
Thick ‘coalesced’ spines 7–9(11) µm
Long spedostump at maturity
Long spines 6–8(10) µm
The slate-morphotype has from subglobose to clavate or pyriform basidiocarps with bright yellow mycelium at the base, light brown spore mass colour and spores 7.5–9(maximum size, 11) µm in diameter, with short spines in light microscopy. Basidiocarps belonging to this morphotype were found on clayey slate-derived soils, most of them near Abenojar (Ciudad Real, Spain). Although the climax vegetation of this region corresponds to a mixed forest of Q. ilex and Pyrus bourgeana with Q. suber, the basidiocarps of Pisolithus tinctorius were frequently associated with Cistus ladanifer, an evergreen shrub that appears after forest disturbance.
The acid-morphotype has typically large basidiocarps (often > 12 cm long) with a massive pseudostump, brown gleba, dark brown spore mass colour and spores 7.5–9(maximum size, 11) µm in diameter with long spines. Basidiocarps showing this morpho-logy were found on acid soils, most of them near Chapinería (Madrid, Spain). This region has acid sandy soils, and the climax vegetation corresponds to forests of Quercus ilex (evergreen oak). In these areas, Pisolithus usually fruits in abandoned lands of cultivation colonized by young evergreen oaks. The specimen esp07, collected in a pine plantation in Kenya (present study), displayed the features of this morphotype, as did the pine morphotype described by Martin et al. (1998) in Kenya (collection no. 5111). A photograph of the habit of this morphotype was reported by Martin et al. (1998).
The basic-morphotype has subglobose basidiocarps with a wide columnar base or pseudostump attached to the substratum by thick basal rhizomorphs. It displayed golden gleba in youth, brown spore mass colour and spores 7–8.5(maximum size, 10) µm in diameter (without ornamentation) with short spines. Samples displaying this morphotype were collected nearby Fuentidueña (Madrid, Spain) on calcareous and marl-gypsum soils, and under Quercus ilex in areas where small evergreen oaks colonize abandoned lands of cultivation. This morphotype was also found in the regions of Murcia, Granada and Valencia on calcareous soils and associated with Q. coccifera and Pinus halepensis.
In spite of not being found in the study region; two other morphotypes were considered, namely the Eucalyptus-type II and Afzelia-type, based on material collected in Kenya. Morphological features for these morphotypes were similar to those reported by Martin et al. (1998) and Cairney et al. (1999). Of these three morphotypes; only the Eucalytus-type I was founded in the Mediterranean Region. These basidiocarps were characterized by clavate to cylindrical (often large, > 12 cm long), and it is characterized by golden gleba, ochraceous spore mass colour and spores 6–7.5(maximum size, 10) µm in diameter with long spines. This morphotype occurred in several plantations of Eucalyptuscamaldulensis, in the provinces of Cáceres and Huelva (Spain), and in Morocco.
ITS sequence variation
A high similarity was found in the 5.8S sequence, confirming that the DNA was sequenced from the target organism, namely Pisolithus, and not from a contaminant (Cullings & Vogler, 1998). By contrast, the ITS1 and ITS2 regions showed significant sequence variation among the different isolates. Sequence comparison showed transitions, transversions and indels. The length variation in amplified ITS/5.8S DNA (data not shown) from the different isolates resulted from deletion and insertions in the ITS regions.
The variations in the length of the ITS sequences were attributable to deletions and insertions, and some gaps were introduced in order to align the sequences. We could define the following regions: a short 3′-partial 18S (position nos 1–16), ITS-1 (from 16 to 203 in alignment), 5.8S (from 204 to 374), ITS-2 (375 to 639), and a short 28S 5′-partial (from position nos 640 to 660 in the alignment). The 5.8S region showed up to approx. 6.6% nucleotide variation, the ITS1 approx. 63%, and the ITS2 region approx. 56%, according to the Jukes and Cantor’s distance index (deletions taken into account).
Pairwise nucleotide-sequence divergences between Pisolithus isolates, calculated using the Junkes and Cantor’s index, ranged from 33% to 0.18%. Pairwise comparison showed five ITS-types, which correlated well with the morphotypes. The variability of ITS/5.8S sequences within a type ranged up to 10%, a value higher to those reported within other basidiomycete species. Distances between types were always over 12.8%. The basic-type and the acid-type were lightly related to each other (Fig. 2). Divergences ranged from 12.7% to 23.95% among types associated to North Temperate hosts, whereas the other types showed 21.2–33.4% nucleotide divergence from each other.
Trees inferred by neighbour-joining (NJ) and parsimony method (PM)
A distance-based tree was obtained using the neighbour-joining method and the Jukes and Cantor’s one-parameter distance (Fig. 2). The 50% majority-rule consensus tree generated by bootstrap heuristic search with 1000 iterations using PAUP 3.1.1 is presented in Fig. 1. The topologies of both distance-based and cladistic trees were congruent, showing three major clusters and clades. Branch stability, assessed by 1000 bootstrap replicates, identified the same well-supported groupings in both neighbour-joining tree (NJ) and parsimony tree (PM). In addition, both trees supported the clear differentiation of the five major groups previously revealed by sequence pairwise comparison.
All Mediterranean host-associated strains nested in a so-called ‘North Temperate grouping’, which was moderately supported by bootstrap analysis (NJ = 81%, PM = 75% of 1000 iterations) (Figs 1, 2). We defined it as ‘North Temperate grouping’, because it only gathers isolates from North Temperates hosts. It displayed three main branchings. The first branch comprised isolates collected on slate-derived soils (slate-type) (NJ = 100%, PM = 100%), which were strikingly separated from the rest of the samples collected on acid soils. The rest of the North Temperate host-associated isolates clustered together (NJ = 97%, MP = 75%) and then divided into two branches, both strongly supported by bootstrap values. The first branch comprised strains occurring on calcareous soils (under Quercus or Pinus halepensis) (basic-type) (NJ = 100%, PM = 100%), whilst the second comprised the acid-type isolates (NJ = 100, PM = 100%). No isolates collected on slate-derived soils were nested within the acid-type grouping, even though derived-slate soils are acid. It is noteworthy that Kenyan isolates under pine (esp, present work; and isolate 5111 (AF003916)), the strain PF (AF143234) from France, and the North American strain Pt-301 (AF143233) collected under pine, were nested within the acid clade.
Isolates associated with Australian hosts grouped into two distinct clades, which corresponded to the ITS type I and type II of Cairney et al. (1999). The Eucalyptus type I clade grouped together isolates from eucalypt plantations in the Mediterranean region (mar01, mar02, ast05, mam17, present work) and Brazil (isolates 444 and Pt-90A) and Australian isolates (PM = 100%). Similar results were obtained by the phenetic analysis (NJ = 100%). In addition, isolates of the Eucalyptus type II and Kenyan isolates under Afzelia, were closely related (MP = 99, NJ = 90). Finally, in spite of being weakly supported by bootstrap analysis, isolates nonassociated with North Temperate hosts, namely Eucalyptus type I, Eucalyptus type II and Afzelia type, clustered and grouped together (NJ = 81%, MP = 75%).
Pisolithus diversity in the Mediterranean region
The defined morphotypes correlate well with the edaphic features of the area where specimens were collected, with clusters obtained by distance methods (Fig. 2) and with clades obtained by parsimony methods (Fig. 1). Furthermore, we observed a lack of restriction fragment polymorphism in the amplified ITS within morphotypes when using AluI, HinfI and MboI enzymes, which contrasted with polymorphism observed between morphotypes (data not shown). From our results, it may be inferred that the soil features are the main ecological factors that may account for the variability of the Pisolithus isolates occurring in the indigenous Mediterranean vegetation. Three main ecotypes can be proposed. The first ecotype (slate-type) includes isolates living on slate-derived soils. Most of the samples belonging to this type were collected near Abenojar (Ciudad Real, Spain). In this region basidiocarps of Pisolithus were often found in scrub of the evergreen shrub Cistus ladanifer, originated as a result of degradation of evergreen and cork oak forests. Indeed, since the slate-type was only found in association with Cistus, this ecotype may be Cistus-specific, as has been shown for other Mediterranean mycorrhizal fungi (i.e. Lactarius cistophilus Bon & Trimbach and Leccinum corsicum (Roll) Singer). The second Mediterranean ecotype of Pisolithus comprises strains living on siliceous soils, associated with Q. ilex and Pinus species. On siliceous soils, basidiocarps of Pisolithus occurred under small evergreen oaks colonizing abandoned lands of cultivation. Lastly, a specific type occurred on basic soils (limey or marl-gypsum soils), associated with evergreen oaks and basophylous pines. In fact, most of the samples collected on calcareous soils were associated with the sclerophilous Q. ilex and Q. coccifera and under Pinus halepensis.
Our sampling within other parts of Europe and North America is limited. Further study will be necessary to state on whether these three lineages are widespread across the Northern Hemisphere. Furthermore, the apparent correlation of each of these clades with soil type needs to be interpreted with caution. It could well be that host, rather than soil, might be what really correlates with these clades. In addition, one of the pitfalls of basidiocarp surveys is that the closest potential host is regarded as host, which is not always true. Further study on Pisolithus host specificity should be based on ectomycorrhizal surveys in order that hosts can may truly determined.
The groupings, obtained by the analysis of the ITS sequences, correlated well with basidiocarp morphology, which suggests that these ecotypes might correspond to different species or species groups. These data suggests that these clades are definitely different from each other. Moreover, at 10% sequence divergence there is a very high likelihood that further species-level differences exist within each of these groups. The basic-type and slate-type clades comprised only Mediterranean host-associated strains, whilst the acid-type included an isolate associated with pine in North America (Pt301) and another pine-associated strain from Kenya (isolate 5111) (Martin et al., 1998). The last ecotype might represent a species complex widely distributed in the Northern Hemisphere and formed by strains associated to different Northern Temperate hosts. In contrast, the slate and the basic-ecotype might comprise species endemic to the Mediterranean region and characteristic of calcareous soils and clayey slate-derived soils, respectively. These data confirm the host specificity grouping of Pisolithus taxa suggested by several other studies. In addition, based on our study we can now report for the first time a likely phenomenon of edaphic vicariance in Pisolithus, at least for the taxa naturally occurring in the Mediterranean region.
Pisolithus occurring in exotic plantations
In general, introduced trees may form mycorrhizas with either indigenous broad range ectomycorrhizal fungi or with introduced host-specific ectomycorrhizal fungi. Our sequence analysis shows that isolates occurring in eucalypt plantations in Mediterranean regions group together with Australian Eucalyptus-type I isolates, suggesting that these isolates were introduced with eucalypt seedlings. These results are similar to those reported by Martin et al. (1998) in Kenya, where these authors suggested the introduction of Pisolithus taxa with exotic Eucalyptus and Pinus. The high degree of similarity between the Australian Eucalyptus-type I isolates and Brazilian collections in eucalypt plantations (GenBank accession nos U62666 and AF140547), indicates that such species were also introduced to Brazil. Since isolates under Eucalyptus from the Northern Hemisphere group together with isolates of Australian origin, migrations of Australian strains must have happened during the last two centuries, as a result of extensive plantations of Eucalyptus. Our data suggest that the Eucalyptus-type I seems to be the main Pisolithus morphotype spread worldwide through eucalypt plantations. However, further study would be required. In fact, the placement of the samples 5110 (accession no. AF003914), collected in eucalypt plantations in Kenya, into the Eucalyptus-type II grouping, suggests that this morphotype has also been introduced to other regions in a similar fashion, as Martin et al. (1998) already indicated for Kenya.
Because Pisolithus tinctoriussensuCoker and Couch (1928) appears to be only associated with North Temperate hosts (e.g. pines and sclerophyllous oaks) it is likely of Holartic origin. This could account for the similarity between isolate esp07 (present work) and the isolate 5510 (Martin et al., 1998), both associated to Pinus caribaea in Kenya, and the isolates widely distributed in the Northern Hemisphere under pines. In a similar fashion, the introduction of extensive plantations of Eucalyptus spp. in other parts of the world may have led to a world-wide dissemination of the Australian Pisolithus; accounting for the grouping of isolates collected under Eucalyptus in Brazil (strain 441), Kenya (strain 5510) and in the Mediterranean (mam17, mar01, mar02 and ast05, present work) with isolates from Australia.
Morphological versus ecological species
Pisolithus has often been regarded as a monospecific genus with a worldwide distribution, able to grow under many soil environmental and varying climatic conditions (Coker & Couch, 1928; Pilát, 1958; Marx, 1977). In contrast, the pairwise distances between Pisolithus ITS sequences (Anderson et al., 1998; Martin et al., 1998, present work), is extensive when compared to those found in some other fungal species (Bruns et al., 1991; Hibbett et al., 1995). Comparative levels of sequence divergence between distinct species of basidiomycetes are interpreted as indicating significant evolutionary divergence. This suggests that we are dealing with a species complex. By using incompatibility tests and basidiospore spine morphology, Kope & Fortin (1990) separated three groups of Pisolithus, and suggested that Pisolithus is composed of several biological species. A number of authors have recently provided support for this hypothesis (Anderson et al., 1998; Martin et al., 1998; Cairney et al., 1999). By using a combined RAPD and ITS sequence approach, Anderson et al. (1998) analysed carpophores collected within a 90 km2 area in Australia, which led them to consider two different species, called type I (CS01, W15, W16, R01) and type II (LJ07 and WM01). Cairney et al. (1999) reported a new ITS-type (type III). Unfortunately sequences of this so-called type-III by Cairney et al. (1999) were not available in the DNA databases, and hence were not included in the present work. Martin et al. (1998) obtained the ITS sequences from three Pisolithus morphotypes collected in Kenya, and compared them with the sequences published by Anderson et al. (1998). They confirmed the presence of at least three distinct Pisolithus genotypes, with < 71% sequence similarity, and thus suggested that these types may belong to separate species. Our data extend these works by introducing an ecological point of view. Since the ecological species concept defines a species as the set of organisms exploiting a single ecological niche or a particular habitat, in the Mediterranean region there may be at least five ecological species of Pisolithus, depending on the host (either exotic or native to the Mediterranean Basin) and the type of soil.
But, should these species be regarded as morphological species? According to Grand (1976), morphological characters should be considered carefully. Spores mass colour and spore ornamentation seems to be dependent on spore maturation. However, in some cases, morphological differences could be of diagnostic value to describe new morphological species, i.e. the Afzelia-type (M. Honrubia, unpublished). Based on the combination of phenetic (culture features and isoenzyme profiles) and genetic markers (RFLP in the ITS regions), and studying Pisolithus isolates from Scandinavia, Europe, North America and the Philippines, Sims et al. (1999) indicated that such isolates probably represent four ‘species groups’ formed by up to 10 Pisolithus species. The integration of studies on typus material preserved in herbaria with molecular data would help to solve the taxonomy of Pisolithus. For instance, the typus specimen of Pisolithus tinctorius was collected under a conifer in North America, and described with brown spore mass with spiny basidiospores (Coker & Couch, 1928), and according to the data obtained, may correspond to the acid-type herein analysed. Other species might be present in the Mediterranean Region, as Calonge & Demoulin (1975) suggested. There might be at least two additional morphological species, which would correspond to the basic-type and the slate-type. In addition, Australian forms of Pisolithus seem to correspond to other species different from Pisolithus tinctorius. The species described by Coker & Couch (1928) might be a North Temperate species, afterwards wide-distributed through other regions due to the exotic plantations of pines. Furthermore, the present molecular data support the idea that Pisolithus tinctorius may not be native to Australia (Bougher & Syme, 1998).
Forestry and environmental implications
The results reported in the present work are of importance for forestry and environmental issues. Selection of indigenous Mediterranean strains of Pisolithus adapted to the soil features where trees are intended to be planted is recommended. The results indicate that co-introduction of Australian Pisolithus could have contributed to the success of eucalypt plantations in the Mediterranean regions. Our results suggest that it seems worthwhile to select strains of Australian origin as inoculants for eucalypt plantations in the Mediterranean region, rather than indigenous isolates.
Burgess et al. (1995) showed that Pinus-associated strains of Pisolithus did not form true ectomycorrhizas on Eucalyptus. Díez et al. (2000) reported that while strains collected in cork oak stands formed ectomycorrhizas with vitroplants raised from somatic embryos, strains collected associated with Eucalyptus did not form mycorrhizas. In the present work, we have not detected either invasions of exotic Pisolithus strains into native Mediterranean forests or out-competition with local mycorrhizal communities. Martin et al. (1998) also described the introduction of Pisolithus taxa associated to exotic plantations of Eucalyptus and Pinus and an apparent lack of transfer of the introduced isolates to indigenous trees. All this might be a result of host specificity and could lead one to think that there is little risk involved with using exotic strains of Pisolithus in Eucalyptus plantations. However, studies based on larger samplings and more loci are required.
We thank Dr F. Martin (INRA-Nancy, France) for suggestions on an earlier version of the manuscript, and to Dr I. Alexander (University of Aberdeen, UK) and two anonymous reviewers for critical comments on the manuscript. We also thank Dr A. Casares (Oviedo University), Mr A. Hernández (Alborada Tree Nursery), the Curator of University Alcalá Herbarium (AH) and Dr Abourouh (DREF-Morocco) for providing some of the isolates. We also appreciate Dr G. Moreno’s taxonomic assistance. This work received financial support from the INIA (SC98-030) and the AECI. J. Díez is grateful to the EU for a postdoctoral fellowship (Contract HPMF-CT-1999–00174). The present study is a joint project between the Plant Biology Departments of the Universities of Alcalá and Murcia and Alborada Tree Nursery.
J. Díez and B. Anta contributed equally to this work.